This application note was prepared in part as a response to recent customer requests for guidance and instruction on how to use ADS to model the small- and large-signal performance of Mini-Circuits Amplifiers based on our published S-Parameter data in .s2p format. In the following sections, we will discuss utilizing how to use .sNp and .txt files to model small- and large-signal performance. We will then demonstrate how to model large-signal performance using both the Amplifier2 and AmplifierS2D behavioral models in ADS, including compression parameters and third-order intermodulation distortion (IP3) in the AmplifierS2D model.
Small Signal S-Parameter Model
Complex S-Parameter data is provided on the Mini-Circuits website for essentially all Mini-Circuits components using a .sNp or .txt file format. These resources allow the customer to use ADS® to model and plot cascaded small signal performance for our components using the .SnP and .txt file formats to model small signal performance Data Item. See Figure 1 below for example.
Figure 1: sNp data item in ADS.
Larger Signal Behavioral Models
ADS can simulate cascaded small and large signal behavior (Pout, Gain, IP3, Harmonics vs. Pin, for example) using Mini-Circuits’ published S-Parameter data, various ADS amplifier behavioral models, and Harmonic Balance simulation.
One technique uses the ADS Amplifier2 behavioral model and a Generalized Multi-dimensional Data (.mdf) file. The .mdf file is easily derived from our published S-Parameter data by changing the header line and adding Begin and End statements. The original .s2p and the newly generated .mdf file formats for our PHA-102+ amplifier are shown below.
Original:
Modified:
…
The variable names freqW, S11dB, S11Ang, S21dB, S21Ang, S12dB, S12Ang and S22dB, S22Ang used in the format line (%) are user defined and arbitrary but must be called out as variables and referenced in the ADS schematic file. The term (real) indicates the entry is a “real” number. Each variable in the format line (%) gives the order for the data in the lines to follow.
The DAC element (Data Access Component) then acquires the data associated with each of these variables. Detailed information regarding the use of .mdf files may be found in the Agilent Technologies Advanced Design System Tutorial, “Using Circuit Simulators”.
The required Variable, S-Parameter and Parameter Sweep blocks are described in Figure 2 below.
Figure 2: Variable, S-Parameter and Parameter Sweep blocks in ADS Amplifier2 behavioral model with .mdf file for PHA-102+ MMIC amplifier.
Variable Definitions
As can be seen from Figure 2 above, the required parameters are pulled from the DAC using Data Item blocks VAR1, 2, 3, 4 and 5.
Parameter Sweep Definitions
The Parameter Sweep block defines the Sweep Variable (SweepVar), which has been arbitrarily defined as “freqW”. It must also specify the S-Parameter or Harmonic Balance Instance Name (SP1 or HB1 as required) referenced in the respective S-Parameter or Harmonic Balance simulation blocks. The Start and Stop frequencies are specified using the sweep variable “freqW” as shown in Figure 3 below. Step Size or points/decade may also be specified using this element.
Figure 3: Parameter Sweep block settings in ADS.
S-Parameter and Harmonic Balance Frequency Definitions
For S-Parameters, the Start and Stop frequencies are specified using the previously defined variable “freqW”. The step size should match that defined in the Parameter Sweep element. For harmonic balance, the fundamental frequency, is given by “freqW”, and the number of harmonics used in the analysis must also be specified.
Small Signal Results (S-PARAMETERS)
The 2-port S-Parameters for the PHA-102+ are obtained by running an S-Parameter simulation and are presented in Figure 4 below.
Figure 4: S-Parameter Simulation of PHA-102+ MMIC amplifier.
These results are, of course, identical to those obtained using the original .s2p file provided by Mini-Circuits.
Large Signal Results (HARMONIC BALANCE): Amplifier2 Behavioral Model
Large signal performance may be modelled using the Amplifier2 behavioral model, as shown in the schematic below (Figure 5), and by performing a Harmonic Balance simulation.
Figure 5: Amplifier2 behavioral model used for large signal performance simulation of PHA-102+.
Simulated, 2-port, large signal parameters (Pout vs Pin and Gain vs, Pin) for the PHA-102+ as a function of the input power variable, Pin, are presented in Figures 6a – 6c below for a 6 GHz carrier frequency. The input power range is specified within the Harmonic Balance block.
Figure 6a: Power sweep parameters specified in Harmonic Balance block.
Figure 6b: Pout vs. Pin
Figure 6c: Output power vs. input power.
Figure 6d: Output power and gain vs. input power.
Identical results may also be obtained by including the amplifiers compression characteristics (Psat, GainCompSat, GainCompPower and GainComp) in the behavioral model (see below) by using a second DAC to read the parameter values directly from a second .mdf file.
It should be noted that different carrier frequencies require that the parameters above either be manually swapped or parametrically read directly from the DAC for each carrier frequency under consideration because ADS does not permit multiple frequencies to be analyzed simultaneously. The Amplifier2 behavioral model allows for the correct parameters to be read from the second DAC simply by sweeping the variable freqW.
Figure 7a: TOI and compression data contained in 2nd DAC element
7b: Schematic for DAC2 element.
Figure 7c: .mdf file for DAC2 element.
By using an alternate form of the Amplifier2 behavioral model, Second and Third Order Intermodulation performance, as well as gain compression and output power saturation characteristics, may be modelled by changing the behavioral model parameters to include SOI and TOI and removing the compression and saturation parameters shown above. An example is given below.
Figure 8a: Specifying SOI and TOI only.
However, ADS restricts the analysis to one or the other (compression and 3rd order intermodulation or 3rd and 2nd order intermodulation, but not both simultaneously) and only at one frequency per simulation.
Figure 8b: Pout and gain vs. Pin for 2nd form of Amplifier2 behavioral model.
Results are similar but it should be clear that tuning the TOI parameter may be necessary to match the measured (published) compression results. For this reason, it is suggested to use the first form of the Amplifier2 behavioral model. While this is preferred, it requires knowledge of the saturated properties of the amplifier which are not always specified but are easily measured or estimated.
Large Signal Results (HARMONIC BALANCE): AmplifierS2D Behavioral Model
Large Signal performance may also be modelled using the AmplifierS2D behavioral model and an .s2d data file. This is similar in format to the .mdf data file used for the Amplifier2 model but includes compression or intermodulation data. For this case, no DAC element is needed. An example schematic is given below.
Figure 9: Example schematic for amplifier large signal performance modelling using the Amplifier S2D behavioral model.
The .s2d file is also easily derived from our published S-Parameter data by changing the header line and adding Begin and End statements. The original .txt and the newly generated .s2d file formats for our PHA-102+ amplifier are shown below.
Original:
Modified:
…
The option line: # AC (Hz S DB R 50 FC 1 0) specifies the units and the format line: % F n11x n11y n21x n21y n12x n12y n22x n22y, gives the order for the data in the lines to follow. It is convenient to associate the variables used in the Amplifier2 model with those of the AmplifierS2D model, but it is not required. The option line is similar to the original except for the FC 1 0 term. The general expression is given by FC m b where Fout = m*Fin + b. In our specific case, m=1 and b=0 indicating a non-frequency translating device.
Compression data is appended after the small signal data as described below. There are seven mutually exclusive formats for expressing large signal response in an S2D file, each corresponding to one type of GCOMP block. We suggest the GCOMP6 format which allows specifying IP3 and P1dB, Psat and Comp_Sat simultaneously. GCOMP6 requires knowledge of the saturated properties of the amplifier which are not always specified but are easily measured or estimated.
Figure 10: GCOMP6 format for expressing large signal response data in the S2D file.
Only one frequency is permitted per simulation. The .s2d file may contain multiple frequency and compression points, but unselected points must be commented out. Changing the analysis frequency requires manually changing the data in the .s2p file. Detailed information regarding the use of .s2d files and the different gain compression models may be found in the Agilent Technologies Advanced Design System Tutorial, “Using Circuit Simulators.”
As before, the variable names freqW and Pin are arbitrary but must be called out as variables and referenced in the ADS schematic file.
Small Signal Results (S-PARAMETERS)
The resulting 2-port S-Parameters for the PHA-102+ are presented in the figure below.
Figure 11: Small signal results using the AmplifierS2D behavioral model.
The small signal results are identical to those using the .mdf and .s2p file formats.
Large Signal results are shown below in Figures 12a and 12 b, and are identical to those obtained using the Amplifier2 model with gain compression data.
Figure 12a: Pin vs. Pout simulation results using the AmplifierS2D behavioral model with compression data appended using the GCOMP block.
Figure 12b: Gain vs. Pout simulation results using the AmplifierS2D behavioral model with compression data appended using the GCOMP block.
Third-Order Intermodulation Distortion (IP3)
As discussed previously,third-order intermodulation distortion may be accounted for with the GCOMP6 AmplifierS2D behavioral model by simply appending the required data after the S-Parameter data. For example, at 6 GHz, a value of +35.32 dBm has been utilized in the model.
The ADS schematic and the Harmonic Balance simulation must be modified to perform the 2-tone IP3 analysis. A representative schematic is given below.
Figure 13: Sample schematic for 2-tone IP3 analysis using Harmonic Balance simulation in the GCOMP6 Amplifier S2D behavioral model.
The single tone source has been replaced by P_nTone, where we have defined two carrier frequencies and their respective power levels. The first frequency, Freq[1], is equal to freqW + deltaF and the second, Freq[2], is freqW-deltaF, where the deltaF variable is the offset from the nominal carrier frequency. For the above we have set deltaF to 1 MHz which results in two carriers separated by 2 MHz. Each power level, P[1] and P[2], are identical and set by variable Pin, which has been defined in the HB1 Harmonic Balance block to vary from -13 to -3 dBm in 1 dB steps.
Simulation results are as follows.
Figure 14: Third-order intermodulation simulation results using the GCOMP6 AmplifierS2D behavioral model.
Using Markers 17 and 18 we may calculate the output IP3 at an input power of -3 dBm to obtain IP3[Pin=-3dBm] =Pout[m1] + (Pout[m1]-Pim3[m2])/2 = +32.5 dBm. Repeating this at an input power of -13 dBm yields IP3[Pin=-13dBm] = + 35.0 dBm, which agrees with the +35.42 (revised) dBm used in the GCOMP6 model and illustrates the effect of increasing input power on IP3. Model results should be compared to measured and adjustments, if necessary, should be applied to the GCOMP6 data entries.
Output IP3 may also be calculated using the IP3out cell and ipo1 function.
Figure 15: ipo1 function.
Results are shown below and agree with the previously derived output IP3 values.
Figure 16: Results from IP3out cell in ipo1 function
Of course, identical output IP3 results are obtained using the Amplifier2 behavioral model as defined earlier. The Amplifier2 with DAC model may be preferred over the AmplifierS2D model because the compression and 3rd order intermodulation data may be simulated at different carrier frequencies without having to adjust .s2d file parameters for each frequency. The variable freqW may simply be changed in the Variable block and the correct data will automatically be selected from the Amplifier2 DAC element.
To better demonstrate the correlation between measured and modelled performance, measured compression and 3rd order intercept point data for the PHA-102+ are presented below. Measured OIP3 was consistent with the original AmplifierS2D behavioral model (published +37.32dBm OIP3) to within a ~2 dB offset. A revised OIP3 value of +35.42dBm (post measurement) easily accounts for the offset as demonstrated by the data below. The excellent tracking between measured and modelled OIP3 data demonstrates the validity of using either Amplifier2 or AmplifierS2D behavioral models.
Figure 17: Measured OPI3 of PHA-102+ vs. original and revised S2D behavioral models.
Bottom Line
There is a good deal of information that can be obtained using the existing S-parameter data that Mini-Circuits publishes for itscomponents. Many design engineers are familiar with these techniques and use them accordingly in their design workflows. For those who may be unfamiliar with the full utility of these resources, this note may provide valuable tools to your toolkit.
You can try these modelling techniques using the provided S-Parameters for all Mini-Circuits amplifiers >
You must be logged in to post a comment.